Circulating tumor DNA positivity identified patients with high-risk muscle-invasive urothelial cancer who were likely to derive improvements in disease-free survival and overall survival from adjuvant atezolizumab vs observation.
Circulating tumor DNA (ctDNA) positivity identified patients with high-risk muscle-invasive urothelial cancer who were likely to derive improvements in disease-free survival (DFS) and overall survival (OS) from adjuvant atezolizumab (Tecentriq) vs observation, according to an analysis of the phase 3 IMvigor010 trial that was presented during the ESMO Immuno-Oncology Virtual Congress 2020.1
In IMvigor010, 809 patients with muscle-invasive urothelial carcinoma at a high risk of recurrence following primary resection were randomized 1:1 to receive atezolizumab or 1 year or observation. Results, which were presented at the 2020 ASCO Virtual Scientific Program, showed that the median DFS with adjuvant atezolizumab was 19.4 months vs 16.6 months with observation, failing to meet the study’s primary end point (HR, 0.89; 95% CI, 0.74-1.08; 2-sided P = .2446).2 OS follow-up is ongoing.
In an exploratory analysis of IMvigor010, investigators prospectively evaluated whether post-cystectomy detectable plasma ctDNA is associated with worse prognosis, whether atezolizumab provides a DFS or OS benefit versus observation in patients with detectable ctDNA, and whether ctDNA clearance in ctDNA–positive patients occurs at a higher rate with atezolizumab vs observation.
Of 581 biomarker-evaluable patients, there were 281 on the observation arm and 300 on atezolizumab. On the observation group, 98 patients (35%) were ctDNA-positive and 183 (65%) were ctDNA-negative; these numbers were 116 patients (39%) and 184 patients (61%) on the atezolizumab arm, respectively.
The DFS outcomes were similar in the ctDNA biomarker-evaluable population compared with the ITT group (HR, 0.88; 95% CI, 0.70-1,11). However, data did show that that ctDNA positivity served as a poor prognosis for DFS (HR, 6.30; 95% CI, 4.45-8.92; P < .0001) and OS (HR, 8.00; 95% CI, 4.92-12.99; P < .0001).
When comparing atezolizumab and observation in ctDNA-positive patients, however, the median DFS was 5.9 months and 25.8 months, respectively (HR, 0.58; 95% CI, 0.43-0.79), and the median OS was 25.8 months and 15.8 months, respectively (HR, 0.59; 95% CI, 0.41-0.86). Results also showed that ctDNA clearance was linked with improved outcomes in the atezolizumab group; moreover, ctDNA-positive patients with high tumor mutational burden (TMB) had improved outcomes vs those with TMB-low patients, as well as those with PD-L1 IC2/3.
The data have practice-changing implications not just for the future of bladder cancer treatment, but other malignancies as well, explained lead study author Thomas Powles, MD, MBBS, MRCP.
“We potentially can select patients for adjuvant therapy, based on a ctDNA biomarker; we have not been able to do that previously in this cancer,” said Powles, director of Barts Cancer Institute. “This is the first randomized trial in any cancer in which we achieved this—I'm fairly certain that's the case with immunotherapy. This may not just be limited to urothelial cancer; this could be effective in other cancers to selecting patients for [intervention with adjuvant treatment] and improving outcomes.”
In an interview with OncLive, Powles discussed this analysis of the IMvigor010 study and its implications for ctDNA as a biomarker in urothelial cancer and beyond.
OncLive: This was a very interesting analysis presented at the ESMO Immunotherapy-Oncology Congress. Please discuss the design of this study.
Powles: We did the IMvigor010 study, [which comprised] a big group of people looking at atezolizumab versus observation post-cystectomy in high-risk, muscle-invasive urothelial cancer. We know that patients with high-risk, muscle-invasive urothelial cancer have about a 50% chance of dying of the cancer after the operation. Therefore, identification of patients who potentially benefit from therapy, and also the identification of those patients with a high risk of the cancer coming back, is really attractive.
The IMvigor010 trial was the first trial to look at immune checkpoint inhibitors in the adjuvant setting. What it showed was that when you compare atezolizumab versus observation in high-risk muscle-invasive urothelial cancer, that wasn't associated with a DFS, or an OS benefit. We can identify a high-risk group of patients; indeed, there was a high mortality [rate] associated with the disease. However, the intervention with atezolizumab didn't help out that population.
The next question we asked was: Did the established biomarkers like PD-L1 and TMB help us identify those patients who benefited from atezolizumab? The answer was, “No, those 2 biomarkers didn't.” Therefore, the next question from there was: Why was that the case? The theory was that actually, if you've had a successful operation, it doesn't matter what your PD-L1 status is because those patients are going to relapse.
Therefore, what we were hoping to do is identify a biomarker associated with minimal residual disease (MRD)—those patients where we know cancer after the operation is still present. That's where ctDNA comes in. Essentially, what we did was we used a technology called Signatera; we do whole-exome sequencing on the tumor. We also look at germline material, and we track 16 mutations identified from the primary tumor, so it's a personalized approach. Then, we do polymerase chain reaction in the plasma, and we can then track those 16 mutations. Over half of the patients had 2 or more mutations of those 16 mutations in the plasma, and they were defined as being ctDNA positive. About half of the patients were ctDNA positive after cystectomy.
What were the findings of the analysis? How is ctDNA positivity a positive correlation to atezolizumab response?
The first thing we showed is, if a patient was CTA positive, they had a really high chance of relapse. If they’re negative, it was actually a pretty low [risk of relapse]. Firstly, that biomarker was really prognostic; it's not perfect in terms of sensitivity and specificity, but ctDNA positivity has about an 80% chance of the cancer coming back.
The second piece of the findings that were really intriguing was if a patient was ctDNA positive, and they received atezolizumab versus observation, the hazard ratio for DFS and death was both about 0.55 and 0.56. What that means is, it's reducing the risk of relapse, and it’s reducing the risk of death by about 40% to 50% in that subpopulation of patients with MRD and are at high risk of the cancer coming back. Then, when you look at the patients who are ctDNA negative, those patients had a much better outcome, very low risk of relapse, and atezolizumab wasn't associated with benefit.
The primary findings of the trial, which were important, is that ctDNA is associated with a high risk of relapse and intervention with atezolizumab in that subgroups seems predictive of response to therapy. The final piece of the work that we did is we looked at sequential ctDNA, both at baseline and measured after 6 weeks of atezolizumab.
We could show that 20% of ctDNA–positive patients cleared their ctDNA. Therefore, atezolizumab has effects on the presence of ctDNA, and, number two, in those patients who did [have clearance], the cDNA did much better with great hazard ratios. We could show that there was there was this sort of pharmacodynamic effect.
Then, the last piece is we correlated this with TMB and PD-L1. We showed that in those patients who were ctDNA positive with high-TMB or high PD-L1, it was associated with better outcomes. That links the immune biomarkers with the immunotherapy in this ctDNA-positive population, and it gives you some biological hypothesis and rationale and it links the biology with the findings. We think it's really important, and it's really exciting for the future, because we can now select patients we know who are at higher risk. These data suggest intervention can help these patients.
What do these findings mean for ctDNA as a predictive biomarker in the adjuvant setting? What are the true implications of this study?
These findings have quite significant implications, and they are quite broad. The broad implications are that for the first time, we haven’t been able to identify a high-risk population, we've identified that an intervention is associated with better outcomes. That has not been done before, with immunotherapy in this space. That has big implications in bladder cancer; we're going to do more trials in bladder cancer.
We are going to do [another study where we are] going to track patients, if they're ctDNA positive, we're going to intervene with atezolizumab vs observation; the observation arm will only get therapy once relapse occurs.
When you look at the data on its face value, you know, it's a study of 800 patients; 600 are in the biomarker analysis—it's actually a pretty robust analysis. You could validate these findings in other ongoing trials. If you're able to do that, what we might need to do is move to a place where actually we can select patients for adjuvant therapy.
The last important implication is the broader implication in that we now moving away from tissue-based biomarkers into circulating biomarkers but not just in urothelial cancer. The future of urothelial cancer, breast cancer, prostate cancer—a broad spectrum of cancers may [have] ctDNA circulating biomarkers. Remember tissue biomarkers have many shortcomings, so this may actually be a one of the first steps towards the future of circulating biomarkers, taking over from tissue-based biomarkers in the past.